DOI: https://doi.org/10.15407/pp2017.02.040

Optimization of auto-tuning of programs using neural networks

А.Yu. Doroshenko, P.A. Ivanenko, O.S. Novak

Abstract


Auto-tuning of programs is a method of self-tuning of internal parameters of the program, affecting its speed, in order to achieve high performance indicators, but it can take a lot of time for testing. In this paper, we propose to improve the method of auto-tuning of programs using neural network algorithms and statistical simulation. The automatic learning of the program model on the results of the "traditional" tuning cycles with the subsequent replacement of some auto-tuner calls with an evaluation from the approximation model allows to significantly accelerate the search for the optimal program variant.

 Problems in programming 2017; 2: 40-47


Keywords


auto-tuning; statistical modeling; automation of software developmen; neural networks

Full Text:

PDF (Ukrainian)

References


Doroshenko A., Ivanenko P., Novak O. Hybrid auto tuning model with the use of a static modelling. Scientific journal "Problems of programming". 2016, N 4. P. 27-32.

Naono K., Teranishi K., Cavazos J., Suda R. Software Automatic Tuning From Concepts to State-of-the-Art Results. Springer; 1st Edition. Edition, 2010. CrossRef

Ivanenko P.A., Doroshenko A.Y., Zhereb K.A. TuningGenie: Auto-Tuning Framework Based on Rewriting Rules. 10th International Conference, ICTERI 2014, Kherson, Ukraine, June 9-12, 2014. Revised Selected Papers. P. 139-158. CrossRef

Ivanenko P.A., Doroshenko A.Y. Method for automated generation of autotuners for parallel applications. Cybernetics and system analisys. 2014. N 3. P. 75-83. CrossRef

Tom M. Mitchell. Machine learning. mcGraw-Hill Science/Engineering/Math., 1997.

Russell, Stuart, Norvig, Peter. Artificial Intelligence: A Modern Approach (2nd edition). Prentice Hall.

Jiawei Han, Micheline Kamber, Jian Pei. Data mining: Cencepts and Techniques, 3rd edition. Morgan Kaufmann, 2011.

Tom Fawcett. An indroduction to ROC analysis. Elsevier B.V. 2005.

Fabian Pedregosa, Gaël Varoquaux, Alexandre Gramfort, Vincent Michel, Bertrand Thirion, Olivier Grisel, Mathieu Blondel, Peter Prettenhofer, Ron Weiss, Vincent Dubourg, Jake Vanderplas, Alexandre Passos, David Cournapeau, Matthieu Brucher, Matthieu Perrot, Édouard Duchesnay. Scikit-learn: Machine learning in Python. 2011. JMLR 12, P. 2825-2830.




DOI: https://doi.org/10.15407/pp2017.02.040

Refbacks

  • There are currently no refbacks.